Tangram is an open source all-in-one machine learning framework that provides libraries for languages including Elixir, Go, JS, PHP,  Python, Ruby, and Rust. The core functionality is implemented in Rust and exposed to each language through native extensions or FFI. In this talk, we will share how you can get started writing your own performance critical code in Rust and how to expose it to a number of different programming languages. Finally, we will share learnings from writing our Gradient Boosted Decision Tree implementation in Rust.